Dextran retention tests have become a standard method for characterizing ultrafiltration (UF) membranes. However, most quantitative analyses of dextran transport have been performed with flat sheet membranes using simple stirred cells in which the transmembrane pressure is nearly uniform across the membrane. The objective of this work was to develop a more fundamental understanding of the factors governing dextran retention tests for large pore size hollow fiber ultrafiltration membranes suitable for use in bioprocessing applications, e. g., the purification of vaccines. Experiments were performed using experimental hollow fiber membranes provided by GE Healthcare. Data were analyzed using a theoretical model that accounts for: (1) the axial variation in transmembrane pressure, and in turn the filtrate flux, due to the pressure drop associated with flow through the fiber lumens, (2) the axial variation in the bulk mass transfer coefficient due to the growth in the concentration polarization boundary layer, and (3) the effects of dextran polarization on the local filtrate flux. The intrinsic sieving coefficients were evaluated using available hydrodynamic models assuming a log-normal pore size distribution. The measured dextran retention coefficients were a strong function of the permeate and feed flow rates. For example, the retention coefficient of a 1000 kDa dextran fraction decreased from R = 0.41 to less than 0.02 as the permeate flow rate was increased from 3 to 18 mL/min at a feed flow rate of 120 mL/min due to concentration polarization in the hollow fiber module. The experimental data were in excellent agreement with model calculations over the entire range of dextran molecular weights and flow conditions. The results provide important insights into the proper design and interpretation of dextran retention tests for hollow fiber ultrafiltration membranes. (C) 2011 Elsevier B. V. All rights reserved.